This invention relates to a person identification apparatus which is used for permitting only specified persons to enter a specific room or for preventing a car from burglary or the like.
Conventional person identification apparatus are generally divided into two types one of which uses the physical information such as fingerprint, retina in the dorsal aspect of eyes, signature, voice and the like, and the other type uses cards such as magnetic cards, IC cards, radio responsive cards and the like.
In the apparatus using the physical information, the fingerprint, retina or the like is read as a pictorial image, and subjected to image processing, and collated with the personal pattern registered beforehand. Particularly, in case of the fingerprint identification, light is applied to a fingerprint so as to form shade by the reflected light, which shade is to be used as an input image. The input image is subjected to the preparation in which the fingerprint is changed into a black and white image of about 128-256 gradations with a resolution of about 512.times.512 picture elements, and after the smoothing and the image enhancement are effected to correct unclear portions, the binarization is carried out. Thereafter, minutiae are used as the characteristics of the fingerprint.
In the apparatus using the voice, since not only the personal information but also the contents of the talk and the feelings are included as the information, it is very difficult to extract only the personal information.
In such conventional person identification apparatus, the physical information such as the fingerprint obtained as the two-dimensional image information is a random pattern so that a very complicated image processing technique is required.
Further, although the identification apparatus using the cards is the generally used person identification apparatus at present, it is disadvantageous in that the person must carry the card and there is a possibility that the card is abused when lost.